On Reducing the Search Complexity for PMI Selection in Codebook-Based 5G NR Systems

To obviate the need for full Channel State Information (CSI) feedback in a Downlink (DL) Multiple Input Multiple Output (MIMO), 5G New Radio (5GNR) employs codebook-based limited feedback techniques. These techniques require the User Equipment (UE) to choose the best possible precoder for the downlink transmission from the set of codebook entries, and the corresponding Precoding Matrix Indicator (PMI) is sent as feedback to the Base Station. Various methods have been proposed in the literature to select the best possible PMI. The Exhaustive Search (ES) method that examines all possible entries of the codebook is the conventional method. Our paper presents a low-complexity PMI selection method that leverages the spatial characteristics of the codebook entries. In this method, the UE approximates the transmit beam to estimate the coarse Angle of Departure (AOD), which is then mapped to PMI using a novel PMI-to-angle mapping lookup table. The size of the codebook is then reduced using this coarse estimate and the final PMI is obtained by searching the reduced codebook. Through this proposed method the number of codebook entries to be searched is significantly reduced, and simulations demonstrate that the proposed algorithm reduces the search complexity by about 94%, even while maintaining the same performance as the conventional exhaustive search scheme.